Surprising SEO Results when you Implement Tedious Work at Scale


Session Description:

It is tempting to always be chasing the latest cool tactic, but over 17 years in the SEO industry, I consistently see boring, tedious and foundational work always making the biggest difference for mid-market and enterprise brands. We are talking mundane stuff like mapping huge lists 404s to correct pages with 301 redirects, or selectively consolidating near duplicates to maximize the performance of the canonicals pages, or writing compelling search snippets at scale. Why? Because few have the patience and time to do this. It takes a lot of work if you have to do it manually. In this talk, I will share client success studies, and cover advanced machine learning techniques and code that we use to scale tedious tasks with great results. You will get password protected Jupiter notebooks prepared exclusively for this event that I won’t share online.

When consolidating duplicate content or recovering links from 404s errors, mapping URLs is a critical but super tedious task you need to complete manually.

I’ve seen online many approaches to try to automatically match duplicate URLs using regular expressions, similarity scores or fuzzy search. Those can work to reduce the amount of manual work, but still, leave a lot of work to complete.

I will cover our approach to using learnable similarity scores using machine learning, deep learning, and small training sets to achieve 95% accuracy.

I plan to share an API, code, and tool the audience can use to leverage this technique.

I will share password-protected access to the relevant tools we use internally in the form of Jupyter notebooks created exclusively for the event.

They will try new high-impact SEO tactics they didn’t consider before because they would take too much work to implement manually.

For example:

i) redirect mapping every single 404 URL with external backlinks, even if we are talking about tens of thousands of links

ii) consolidating near duplicate content opportunistically (when there is no keyword loss in the duplicate cluster)

iii) writing unique, compelling and high performing search snippets at scale by leveraging user review copy

You will get an in-depth look at the algorithms, ideas, and frameworks used to research, experiment and automate tedious but valuable work

You will get Jupyter notebooks with relevant code and use cases prepared exclusively for this talk and not share anywhere else.

Targeted – Technical staff (developers, technical SEOs, data scientists) at brands and agencies

Date: November 7, 2019 00:00

Track:

Speaker(s):